Method and apparatus for processing an image of a road to identify a region of the image which represents an unoccupied area of the road

Abstract

A method of processing an image of a scene including a road acquired by a vehicle-mounted camera to generate boundary data indicative of a boundary of an image region which represents an unoccupied area of the road, comprising: generating an LL sub-band image of an Nth level of an (N+1)-level discrete wavelet transform, DWT, decomposition of the image by iteratively low-pass filtering and down-sampling the image N times, where N is an integer equal to or greater than one; generating a sub-band image of an (N+1)th level by high-pass filtering the LL sub-band image of the Nth level, and down-sampling a result of the high-pass filtering, such that the sub-band image of the (N+1)th level has a pixel region having substantially equal pixel values representing the unoccupied area of the road in the image; and generating the boundary data by determining a boundary of the pixel region.

The invention claimed is: 
     
       1. A method of processing an image of a scene including a road that has been acquired by a camera mounted on a vehicle to generate boundary data indicative of a boundary of a region of the image which represents an unoccupied area of the road, the method comprising:
 generating an LL sub-band image of an N th  level of an (N+1)-level discrete wavelet transform, DWT, decomposition of the image by performing an iterative process of iteratively low-pass filtering and down-sampling the image N times, where N is an integer equal to or greater than one; 
 generating a sub-band image of an (N+1) th  level of the (N+1) level DWT decomposition of the image by high-pass filtering the LL sub-band image of the N th  level, and down-sampling a result of the high-pass filtering, such that the sub-band image of the (N+1) th  level has a region of pixels which represents the unoccupied area of the road in the image; and 
 generating the boundary data by determining a boundary of the region of pixels, 
 wherein the boundary data is generated by determining the boundary of the region of the sub-band image by: 
 determining, for each of a plurality of columns of pixels in the sub-band image, a pixel location of a pixel in the column at which a difference between a pixel value of the pixel and a pixel value of an adjacent pixel in the column exceeds a predetermined threshold; and 
 defining line segments in the sub-band image using the determined pixel locations, the line segments defining the boundary of region of the sub-band image. 
 
     
     
       2. The method according to  claim 1 , wherein a first low-pass filter having a first sequence of filter coefficients that are symmetrical is used in at least one iteration of the iterative process. 
     
     
       3. The method according to  claim 2 , wherein the filter coefficients in the first sequence of filter coefficients are set to values in a row of Pascal's triangle having the same number of values as an order of the first low-pass filter. 
     
     
       4. The method according to  claim 1 , wherein the high-pass filtering used to generate the sub-band image of the (N+1) th  level comprises applying a high-pass filter having a second sequence of filter coefficients that are symmetrical. 
     
     
       5. The method according to  claim 4 , wherein:
 alternate filter coefficients in the second sequence of filter coefficients are set to correspondingly located values in a row of Pascal's triangle having the same number of values as an order of the high-pass filter; and 
 each remaining filter coefficient in the second sequence of filter coefficients is set to a value obtained by multiplying a correspondingly located value in the row of Pascal's triangle by −1. 
 
     
     
       6. The method according to  claim 1 , wherein the sub-band image of the (N+1) th  level of the (N+1)-level DWT decomposition is one of an LH sub-band image, an HL sub-band image, and an HH sub-band image. 
     
     
       7. The method according to  claim 1 , wherein generating the sub-band image of the (N+1) th  level of the (N+1)-level DWT decomposition of the image comprises generating an LH sub-band image of the (N+1) th  level of the (N+1)-level DWT decomposition of the image by one of:
 a first process comprising:
 generating a low-pass filtered LL sub-band image by applying a row kernel which defines a low-pass filter across the rows of the LL sub-band image of the N th  level of the (N+1)-level DWT decomposition of the image; 
 down-sampling the columns of the low-pass filtered LL sub-band image by a factor of two to generate a down-sampled sub-band image; 
 generating a high-pass filtered LL sub-band image by applying a column kernel which defines a high-pass filter across the columns of the down-sampled sub-band image; and 
 down-sampling the rows of the high-pass filtered LL sub-band image by a factor of two to generate the LH sub-band image of the (N+1) th  level; 
 
 a second process comprising:
 generating a high-pass filtered LL sub-band image by applying a column kernel which defines a high-pass filter across the columns of the LL sub-band image of the N th  level; 
 down-sampling the rows of the high-pass filtered LL sub-band image by a factor of two to generate a down-sampled sub-band image; 
 generating a low-pass filtered sub-band image by applying a row kernel which defines a low-pass filter across the rows of the down-sampled sub-band image of the N th  level; and 
 down-sampling the columns of the low-pass filtered sub-band image by a factor of two to generate the LH sub-band image of the (N+1) th  level; and 
 
 a third process comprising:
 generating a filtered sub-band image by applying a two-dimensional kernel across the LL sub-band image of the N th  level, the two-dimensional kernel being separable into a product of a row kernel and a column kernel, the row-kernel defining a low-pass filter and the column kernel defining a high pass filter; and 
 down-sampling rows and the columns of the filtered sub-band image by a factor of two. 
 
 
     
     
       8. The method according to  claim 1 , wherein no more than two sub-band images are generated in each level of the (N+1)-level DWT decomposition of the image up to the N th  level, and wherein only the sub-band image is generated at the (N+1) th  level of the (N+1)-level DWT decomposition of the image. 
     
     
       9. The method according to  claim 1 , further comprising setting N to a value that is based on an image resolution of the image, by using a predetermined mapping between values of the image resolution and values of N. 
     
     
       10. The method according to  claim 1 , wherein generating the boundary data comprises determining the boundary of the region of the sub-band image by:
 determining a pixel location of a pixel of the sub-band image of the (N+1) th  level whose pixel value falls below a predetermined threshold; and 
 executing a contour tracing algorithm using the determined pixel location to identify the boundary of the region of the sub-band image of the (N+1) th  level, wherein the boundary separates pixels of the region which are adjacent to the boundary and have pixel values below the predetermined threshold, from pixels outside the boundary which are adjacent to the boundary and have pixel values above the predetermined threshold. 
 
     
     
       11. The method according to  claim 1 , wherein N is an integer equal to or greater than two, and the method further comprises generating refined boundary data, which is indicative of the boundary of the region of the image which represents the unoccupied area of the road, by:
 high-pass filtering an LL sub-band image of a P th  level of the (N+1)-level DWT decomposition of the image and down-sampling a result of the high-pass filtering of the LL sub-band image of the P th  level to generate a second sub-band image which is a sub-band image of a (P+1) th  level of the (N+1)-level DWT decomposition of the image, where P is an integer smaller than N; 
 determining a second boundary, in the second sub-band image, by up-scaling the determined boundary of the region of the sub-band image of the (N+1) th  level using a scaling factor of 2 T , where T=N−P, to generate an up-scaled boundary, and mapping the up-scaled boundary to the second sub-band image; and 
 processing the second sub-band image using the determined second boundary to generate, as the refined boundary data, data indicative of a boundary of a second region in the second sub-band image which contains the second boundary and represents the unoccupied area of the road in the image. 
 
     
     
       12. The method according to  claim 1 , wherein N is an integer equal to or greater than two, the road in the scene has a road marker, and the method further comprises generating boundary data indicative of a boundary of a road marker region of the image which represents the road marker, by
 high-pass filtering an LL sub-band image of a M th  level of the (N+1)-level DWT decomposition of the image and down-sampling a result of the high-pass filtering of the LL sub-band image of the M th  level to generate a sub-band image of a (M+1) th  level of the DWT decomposition, M being an integer less than N; 
 determining a search area in the sub-band image of the (M+1) th  level by up-scaling the boundary determined from the sub-band image of the (N+1) th  level by a factor of 2 D  for D=N−M, and mapping the up-scaled boundary to the sub-band image of the (M+1) th  level, wherein the search area lies within the up-scaled boundary that is mapped onto the sub-band image of the (M+1) th  level; and 
 determining, within the search area in the sub-band image of the (M+1) th  level, a boundary of a second region of pixels of the sub-band image of the (M+1) th  level, the boundary of the second region being indicative of the boundary of the road marker region in the image. 
 
     
     
       13. A non-transitory computer-readable medium storing instructions that, when executed by a processor, cause the processor to perform a method according to  claim 1 . 
     
     
       14. An apparatus for processing an image of a scene including a road that has been acquired by a vehicle-mounted camera to generate boundary data indicative of a boundary of a region of the image which represents an unoccupied area of the road, the apparatus comprising:
 a processor; and 
 a memory, including instructions stored thereon, which when executed by the processor cause the apparatus to:
 generate an LL sub-band image of an N th  level of an (N+1)-level discrete wavelet transform, DWT, decomposition of the image by iteratively low-pass filtering and down-sampling the image N times, where N is an integer equal to or greater than one; 
 generate a sub-band image of an (N+1) th  level of the (N+1) level DWT decomposition of the image by high-pass filtering the LL sub-band image of the N th  level, and down-sampling a result of the high-pass filtering, such that the sub-band image of the (N+1) th  level has a region of pixels which represents the unoccupied area of the road in the image; and 
 generate the boundary data by determining a boundary of the region of pixels, 
 
 wherein the boundary data is generated by determining the boundary of the region of the sub-band image by: 
 determining, for each of a plurality of columns of pixels in the sub-band image, a pixel location of a pixel in the column at which a difference between a pixel value of the pixel and a pixel value of an adjacent pixel in the column exceeds a predetermined threshold; and 
 
       defining line segments in the sub-band image using the determined pixel locations, the line segments defining the boundary of region of the sub-band image.